Differences in search strategies and output between Cochrane Reviews and matching Health Technology Assessment (HTA) report
2HTA Syd, Skåne University Hospital, Lund, Sweden
Creating and running search strategies are complex tasks relying on information specialists working in collaboration with clinical experts, as well as other available resources. Differences in search strategies for the same research question between review teams are common. Where these differences emerge and the impact of different search strategies needs further attention in order to compare and evaluate current search strategy practices.
Objectives: Our aim was to investigate if and how search strategies differ between Cochrane Reviews and HTA Reports on the same research question.
Methods: We identified and compared the search strategies in HTA-reports published by the HTA unit in southern Sweden and matching Cochrane Reviews. When discrepancies in the PICO-S or timing of the search was identified, for example HTA-reports including additional study designs or conducted at different times, the HTA-search strategy was modified to get more comparable search output. In this pilot project we compared the following variables in the HTA-report and the corresponding Cochrane Review: 1. Search terms 2. Subject terms (e.g. MeSH-terms) 3. Limiters/filters (e.g. animal, children) 4. Language restrictions (e.g. no restrictions, English) 5. Study types (e.g. RCT’s, protocols, grey literature and conference abstracts) 6. Number of databases searched 7. Number of hits
Results: Our preliminary results for the initial two review research questions with matching HTA-report and Cochrane review indicates that the Cochrane reviews has more overall hits and included studies, included more databases, used less language restrictions and included more search terms and combinations while the HTA-report included more study types. Different search-limiters were used. Table 1.
Conclusions: There are several differences in search strategies between HTA-reports and Cochrane reviews which has an impact on the number of hits and possibly on included studies. Differences can be due to available resources for screening and the time-to-completion of reviews and reports. Patient, public and/or healthcare consumer involvement None, due to the technical topic of the project.